Comparative Analysis for Email Spam Detection Using Machine Learning Algorithms

Abstract

AbstractThe prominence of the use of communication over the Internet is increasing progressively. Being economical, faster, and easy user interface, the number of email users is increasing tremendously. These led to the gradually increasing activity of spam. Spam emails are unrequested and unimportant emails in bulk. Due to this, there arise major Internet and email security issues that also include a problem of electronic storing space and waste of time. Thus, the identification of spam emails is very necessary. In this paper, four supervised machine learning algorithms, which are Naïve Bayes, support vector machine (SVM), logistic regression, and random forest classifier, are proposed for spam and ham emails classification. Experiments using these four algorithms are performed on prepared feature sets on two different datasets to select the best model with the highest accuracy and less overfitting or underfitting for spam detection. To automate the workflow of building the model and its evaluation, a machine learning pipeline is used in this project. Experimental results show that the overall accuracy of the random forest classifier model is the highest and also has less complexity.KeywordsEmailsSpam emailSpam email identificationMachine learningNaïve BayesSupport vector machineLogistic regressionRandom forest classifier